Breast cancer evolves in a dynamic microenvironment. The epithelial cancer cells subvert and recruit other host cells as the cancer progresses to the invasive stage, in turn, fibrobtasts, myoepithetial cells and the endothelium act upon the cancer cells themselves. We now know that extracelUar matrix (ECM) and growth factors can radically alter the behavior of cancer cells/1, and that context can strongly influence how cancer cells respond to therapeutic agents/2. In the tumor context, the extracellular matrix is both a product of the breast cancer cells and the host cells, and that variation in the tumor cell genome and transcriptome modulate this response. We hypothesize that the response of breast cancer to therapeutic agents is modified by this complex microenvironment. We predict that the cellular composition of the microenvironment is crucial for predicting which agents will successfully control each given breast cancer. We also believe that the failure to consider the role of the microenvironment may partially account for our failure to successfully treat some breast cancers. We therefore posit that it is important to test new therapeutics considering the context of normal tissue and the particular malignant tumor. Further, we postulate that the therapy in turn will modify both the cancer cells and the microenvironment. We have shown for example that ionizing radiation, widely used in breast cancer therapy, persistently alters both the composition of the stroma in mice/3 and cell-cell and ceIl-ECM interaction in human breast epithelial cells in three-dimensional cultures (3D)/4. As a result of fractionated therapy, treated tumor tissue is the norm at the star of therapy. It is thus important to be mindful of the fact that the phenotype of the tumors is constantly changing as a function of the stage and duration, as well as modality of therapy. The objective of this proposal is to critically examine the influence of microenvironmental interactions (i.e. cell-ECM, cell-cell- myoepithelial and stromal- interactions) on therapeutic responses and to use this information to increase the accuracy with which we can model the signaling pathways both experimentally and computationally. We will accomplish this by studying representative subclasses of the 60 breast cancer cell lines that comprise the "system" that is the overall focus of this Integrative Cancer Biology Project. Specifically, we will: Aim 1. Develop 3D cultures of interactions between breast cancer ceils and fibroblasts and/or myoepithelial cells that originate from either normal or tumor tissue and define the breast cancer cell Pathway Logic starting state as a function of co-cultures. Our goal in this aim is to redefine the Pathway Logic model initial conditions so they match those in co-cultures. The revised model can then be used to predict responses to Raf-MEK-ERK pathway manipulations in co-cultures that may be different from those observed in analyses of 2D and 3D cultures in Project 2. Aim 2. Determine the response of breast cancer ceils to Raf-MEK-ERK module inhibitors in the context of specific microenvironments developed in Aim 1, by measuring growth response, gene expression, and phenotype in vitro. These studies will test the Pathway Logic model predictions made based on information from Aim 1. Response information will be used to further refine the model. Aim 3. Develop in vivo models of humanized stroma to test agents that have microenvironment? based signatures of response as defined in Aim 2. These studies will provide Pathway Logic model initial conditions that are consistent with those found in vivo. Pathway Logic model predictions of in vivo responses to Raf-MEK-ERK perturbations will then be tested experimentally.